Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=ysta20
STAR: Science & Technology of Archaeological Research
ISSN: (Print) 2054-8923 (Online) Journal homepage: https://www.tandfonline.com/loi/ysta20
Antimony as a raw material in ancient metal and
glass making: provenancing Georgian LBA metallic
Sb by isotope analysis
Sarah Dillis, Alicia Van Ham-Meert, Peter Leeming, Andrew Shortland, Gela
Gobejishvili, Mikheil Abramishvili & Patrick Degryse
To cite this article: Sarah Dillis, Alicia Van Ham-Meert, Peter Leeming, Andrew Shortland, Gela Gobejishvili, Mikheil Abramishvili & Patrick Degryse (2019): Antimony as a raw material in ancient metal and glass making: provenancing Georgian LBA metallic Sb by isotope analysis, STAR: Science & Technology of Archaeological Research, DOI: 10.1080/20548923.2019.1681138 To link to this article: https://doi.org/10.1080/20548923.2019.1681138
© 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
View supplementary material
Published online: 13 Nov 2019. Submit your article to this journal
Article views: 345 View related articles
Antimony as a raw material in ancient metal and glass making: provenancing
Georgian LBA metallic Sb by isotope analysis
Sarah Dillisa, Alicia Van Ham-Meert a,b, Peter Leemingc, Andrew Shortlandd, Gela Gobejishvilie, Mikheil Abramishvilieand Patrick Degrysea,f
a
Earth and Environmental Science, Centre for Archaeological Science, KU Leuven, Heverlee, Belgium;bAnalytical, Environmental and GeoChemistry, VUB, Brussels, Belgium;cUniversity of Exeter, Exeter, UK;dCranfield Forensic Institute, Cranfield University, Defence Academy of the United Kingdom, Swindon, UK;eGeorgian National Museum, Tbilisi, Georgia;fDepartment of Archaeological Sciences, Faculty of Archaeology, Leiden University, Leiden, Netherlands
ABSTRACT
Sb was frequently used as a raw material, both in ancient glass-making (as an opacifier and decolouriser) and metallurgy (either as an alloying element or as a pure metal). Despite this ubiquity, antimony production has only occasionally been studied and questions concerning its provenance are still not satisfactorily answered. This study evaluates the suitability of Sb isotope analysis for provenance determination purposes, as experiments under lab conditions have revealed fractionation occurring during redox processes in oxidising stibnites and in making opacified glasses. The results of this paper help to evaluate the possible influence of the pyrotechnological processes on the antimony isotope composition of glass artefacts. This paper focuses on the Caucasus as case study by applying mineralogical, geochemical and isotopic analysis to Georgian ores (mainly from the Racha-Lechkumi district) and Late Bronze Age (LBA; 15th–10th century BCE) metallic Sb objects found at the sites of Brili and Chalpiragorebi.
ARTICLE HISTORY
Received 5 December 2018 Accepted 10 October 2019
KEYWORDS
Antimony; provenance studies; glass; metals; Late Bronze Age; Caucasus
Introduction
Sb in the archaeometallurgical record
Finds of antimony in the archaeological record show a long history of use. First, in metallurgy, Sb appears mostly as antimonial copper (see Figure 1). In the southern Levant, Cu-based metals with variable levels of As and/or Sb, ranging in content between 1 and 20%, and traces of Ag, Bi and Ni, appear during the Chalcolithic period (5500–3300 BCE) (Shalev and
Northover 1993; Carmi, Epstein, and Segal 1995; Hauptmann and Gambaschidze 2001; Golden 2014). In Central Italy, these kind of Cu-As-Sb alloys also appear from the early or mid-4th millennium BCE (Chalcolithic period) at various sites (Dolfini 2010). The objects, archaeologically assigned to the Rinaldone culture, are similar in composition to the Levantine Sb-rich alloys with 1–20% Sb and/or As and traces of Ag, Bi & Ni (De Marinis2006; Grazzi et al.2012; Petitti, Persiani, and Pallecchi 2012). In the region of the
© 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
CONTACT Sarah Dillis Sarah.dillis@kuleuven.be Earth and Environmental Science, Centre for Archaeological Science, KU Leuven, Celestijnenlaan 200E, B-3001 Heverlee, Belgium
Supplemental data for this article can be accessed athttps://doi.org/10.1080/20548923.2019.1681138
Caucasus, at the end of the Early Bronze Age (EBA; 3300–2100 BCE) or definitely from the Middle Bronze Age (MBA; 2100–1550 BCE) onwards, Cu-Sb alloys with 1–20% Sb and sometimes with a similarly high As content are found (Hauptmann and Gambaschidze
2001). They are mainly assigned to the Caucasian cul-tures of Bedeni (2500–2200 BCE), Sachkhere (2600– 2000 BCE) and Trialeti (2000–1700 BCE) (Chernykh
1992) and characterised by traces of S (Pike2002). In this aspect they differ from the previously mentioned Chalcolithic Levantine and Italian Sb-rich alloys. During the LBA (1500–1150 BCE) the use of Sb-rich alloys became even more widespread in the Caucasian region, especially in the northern Koban area (Pike
2002). New complex alloys such as Cu-Sb-Pb and Pb-Sb are also found (Meliksetian and Pernicka
2003). At Hasanlu, Iran, Sb-rich alloys have been dis-covered, securely dated to Iron II or Hasanlu IVB (around 800 BCE) (Fleming, Nash, and Swann2011). In the case of metallic Sb objects, the same region in Central Italy during the Chalcolithic period is of inter-est (see Figure 1). At the sites of Ponte San Pietro (Tombs 21 and 22) (Dolfini 2014) , Selvicciola (Tombs 21, 23, and 35) (Grazzi et al.2012; Petitti, Per-siani, and Pallecchi2012), Grotta del Fontino (Zanini
2002) and Montebradoni (Cambi and Cremascoli
1957) multiple Sb beads and buttons have been
found. These pieces of evidence combined with the finds of antimonial copper consistently indicate that the working of antimony began in this region during the Early Chalcolithic period (Pallecchi, Pecchioli, and Tocci 2002; Dolfini 2010). In Mesopotamia, Sb objects are reported from the 3rd millennium BCE (e.g. at Tello, Tell Leilan, Jerablus Tahtani) (Shortland
2002; Moorey1999), and later from the 1st millennium BCE in Egypt at Lahun (Petrie 1891), in Anatolia at Yoncatepe (Belli and Konyar 2001), and in Iran at Hasanlu (Fleming, Nash, and Swann2011). In the Cau-casus, metallic Sb objects are abundant. Theirfirst use-age is identified in the middle of the 3rd millennium BCE, becomes even more prominent during the 2nd millennium and runs far into the LBA and Early Iron Age (EIA; 1200–1000 BCE). Examples are known from sites in Armenia, e.g. at the MBA and LBA levels of the Lchashen and Artik settlements, and in burial contexts at Metsamor (LBA–EIA) (Meliksetian and Pernicka 2003), Redkin Lager (LBA) (Lindsay and Smith 2006), Chesmanis (LBA–EIA), Chambarak
(ninth century BCE) (Tite 2002; Meliksetian et al.
2011), Bjini (EIA–MIA) (Meliksetian et al. 2011), in Georgia in burial contexts at Brili, Treli, Kvasatali, Tsaghvli, Chalpiragorebi, Kvatskhelebi, etc. dating between the 16thand 12th centuries BCE (Hauptmann and Gambaschidze 2001), and in Dagestan small Sb
ornaments were also cast in the Kayakent–Khorochoi culture (2nd millennium BCE) at Mamai-Koutan, Kor-ochoi (Burial 2) and Kayakent (Burial 16) (Chernykh
1992). Interestingly, there is archaeological evidence for Sb mining in the Racha–Lechkumi area (Georgia) at the Sagebi and Zopkhito mines, possibly from the 19th century BCE but definitely from the 17th century onwards (Chernykh1992).
Sb as opacifier and decolouriser in glass production
Sb was not only used in metallurgy, but was also widely used in glass production, either as an opacifier (Sb con-tent usually >1%) in coloured glass from the LBA onwards, or as decolouriser (around 0.5% Sb) from the Hellenistic period (fourth century BCE) onwards.
As far as opacifying agents are concerned, the early history of glass mainly sees the use of calcium and lead antimonates for the production of opaque white (Ca
2-Sb2O7or CaSb2O6) and yellow (Pb2Sb2O7) glasses. A
process of making a precursor is proposed for the LBA Egyptian yellow glasses (Shortland 2002). Apart from calcium and lead antimonate, which have been found to be the predominant opacifiers used in the LBA glasses, sodium antimonite crystals (brizziite, NaSbO3) have been reported associated with lead
anti-monate and calcium antianti-monate as the white opacifier used in the Iron Age glasses from Hasanlu (Stapleton
2011). These Sb-based opacifiers continued to be
used in the Roman world (Shortland2012). Similarily to the LBA yellow glasses, it was suggested that roasted stibnite (i.e. antimony oxide), rather than calcium anti-monate, was added to Roman colourless and blue glass to make these glasses opaque white and opaque blue (Freestone and Stapleton2015). For the Roman yellow and green opaque glasses antimonial litharge or a pre-cursor lead-antimony-silicate (“anime”) was proposed as the antimony source added to the soda-lime-silica glasses (Mass, Wypyski, and Stone 2002; Freestone and Stapleton 2015). These antimonates are not only present in the earliest glasses, Sb opacified-coloured glass is also by far the most common glass used, accounting for at least 90% by weight (Shortland2012). Since small amounts of Fe2O3are usually
uninten-tionally present as impurities in the sand raw material which results in a green–blue tint of the glass, a deco-louriser is needed to obtain colourless glass. The colour can be neutralised by the addition of Mn or Sb by the oxidation of the Fe2+ to the practically colourless or weak yellowish Fe3+ (Sayre 1963; Schreurs and Brill
1984). Antimony levels in sand are typically at the ppm (or sub-ppm) level (Brems 2012; Brems and Degryse 2014), and greater concentrations in glass indicate the deliberate addition of an antimony-bearing ingredient (Freestone and Stapleton 2015). The tech-nology of use and provenance of the raw materials of
antimonates are therefore vital to the understanding of the history of glass.
Objectives
Despite the long-term and widespread use of anti-mony, many questions remain concerning its origin. What is the primary origin of the Sb raw materials used? How was this Sb distributed, by processes of trade, exchange or tribute? What was the nature of its adoption in several technological processes?
For the Levantine antimonial copper alloys, an ore of the tetrahedrite-tennantite solid solution (s.s.) (in roughly equal proportions) with accessory nickel min-erals is suggested by Tadmor et al. (1995). Depositis with such a signature in the Caucasus are likely sources (Tadmor et al.1995). One likely origin of the antimony used as an opacifier in ancient glass making throughout the Near East also lies in the Caucasus (Shortland
2002). Hence, this paper aims at provenancing the Sb raw materials used in the early Caucasian Sb metal-lurgy industry, and investigate possible links to glass making. Metallic Sb beads from the Racha-Lechkumi region (Georgia) are investigated using mineralogical, geochemical and isotopic techniques to have a better understanding of the technology and provenance of the Sb raw materials used in the Caucasian LBA–EIA metal production.
takes place due to a change in oxidation state during digestion of the samples in laboratory study (hypoth-esis 1) and whether fractionation occurs in redox reac-tions during pyrotechnological processing in ancient metallurgy and glass making (hypothesis 2). Only thereafter Sb isotope ratio analysis comparing ores to early metals and glass can be performed.
Materials and methods
Archaeological & geological context
A collection of Sb-rich ores (n = 90) from various regions across the world was analysed with XRD and ICP-OES (n = 42) to obtain mineralogical and chemical characteristics. Furthermore, under the auspices of the Georgian National Museum in Tbilisi, 27 samples of Sb-rich materials were obtained for mineralogical, geo-chemical and isotopic analysis: 9 ores from the Racha-Lechkumi region of which 8 stibnites and 1 tetrahe-drite-tennantite s.s., and 18 metallic Sb objects from the LBA cemeteries of Brili and Chalpirogrebi. The list of samples studied can be found in Appendix I.
The Zemo (Upper) Racha ridge, situated in the dis-trict of Racha-Lechkumi in the Republic of Georgia, is rich in mineral deposits. Geologically, the Racha-Lech-kumi region is centred in an area of Jurassic rocks. It occurs as a narrow slate belt that follows the general WNW-strike of the Greater Caucasus. The slates are distinguished by anomalously high concentrations of As, Sb, Cu, Zn and Au, likely defining a broad region of hydrothermal alteration. In addition, narrow hor-izons of diagenetic pyrite are also present and this
early pyrite is enriched in As, Au and Sb, which could be a source for some of the younger epigenetic metal enrichments (Kekelia et al.2008). The Sb depos-its of Sagebi (W–Sb–Au), Kairobi (Mo–Sb), Sanarts-khia (Sb–Au), Kvardzakehti (Sb–Au), Zopkhito (Au– Sb) and to the west, the Guli Sb deposit, are situated in this district (https://www.mindat.org/min-3782. html, last visited on 20/06/2018).
The region also contains a dense concentration of prehistoric sites associated with early metallurgy. About a hundred locations have been discovered at which copper, as well as antimony and arsenic ores were worked (Chernykh1992). There are also several major deposits of antimony which were exploited in prehistoric times, including: Gona, Zopkhito and Sagebi (Inanishvili, Maisuradze, and Gobejishvili
2010). During the 10–12th of September 2016 an
expedition was carried out by P. Leeming and the Georgian National Museum (GNM) to collect some of this Rachan material. During the expedition the site“Chkornali 1” was visited (N 42° 49′ 21.6′′; E 43° 32′ 51.5′′; 2429 m elevation). Although the mine itself is flooded, three samples were collected from spoil between 5 and 15 m from the entrance. The source of the samples is shown inFigure 2.“Georgia 1” was col-lected close to the Zopkhito Sb-deposit, where traces of antimony smelting have been found near the ore body (Chernykh1992). Samples Georgia 2–5 also come from
ancient mines (sloping galleries and pit faces) in this part of the Racha-Lechkumi district.
Fifteen beads found at the site of Brili, excavated from Grave 32 (11 beads studied), Grave 33 (1 bead) and Grave 1 (3 beads studied) were analysed. The
beads of Grave 32 are part of three different necklaces (see Appendix I): necklaces A and B are bi-lobed beads consisting of two spherical bodies that share a common flat bottom (typical grave goods in the tombs of Brili, Nuli and Kvasatali (Shida Kartli)) (Hauptmann and Gambaschidze2001); necklace C is tri-lobed, consist-ing of three spherical bodies.
Brili is a multi-period cemetery in the region of Racha-Lechkumi, situated to the west of the area of Gona. The site was excavated in nine seasons between 1939–1961 under the leadership of Germane Gobejish-vili (GobejishGobejish-vili1977, 1952). Even though the ceme-tery has not been completely excavated, more than 200 graves have been opened. Local communities used the cemetery intermittently as a burial ground for about 2000 years from the early 2nd millennium BCE (MBA) to the fourth century CE. They are charac-terised by a diversity of belief systems and mortuary practices– earthen pit graves, stone cists, and crema-tion platforms (Sagona 2017). The site is only dated by relative chronologies and is being analysed burial by burial with thus far limited publication (e.g. Pants-khava, Maisuradze, and Gobejishvili1939). This lack of absolute dating is common in the Bronze Age of Geor-gia as a whole, where programmes of14C dating are still in their infancy.
From the Chalpiragorebi cemetery in Tbilisi, in use during multiple periods, spanning the 2nd and 1st mil-lennium BCE, 5 beads fromfive different graves were analysed:“CHPG 1” from Grave 15 belongs to a com-posite spacer-plate necklace of over 80 beads, of which an axe-head shaped bead (like a so-called Central Transcaucasian battle axe) was analysed; “CHPG 2” from Grave 75 is a flattened truncated bicone or lozenge bead; “CHPG 4” are broken fragments from a necklace found in Grave 3; the accession numbers of the two other samples “CHPG 5” and “CHPG 6” are currently being investigated and the site is being prepared for full publication by one of the authors, M. Abramishvili. Unfortunately, not much contextual information in English has been published so far.
Fractionation experiments
The outline of the fractionation experiments can be found inTable 1.
Fractionation experiment 1
Originally, Sb isotope analysis of low Sb content materials was developed by Lobo et al. (2012) isolat-ing Sb usisolat-ing a combination of cation exchange chromatography with Dowex resin and anion exchange chromatography using Amberlite resin (Lobo et al. 2012, 2013, 2014; Degryse et al. 2015). This procedure was altered for high Sb content materials due to the creation of viscous gels instead of clear solutions when preparing such samples. Thefirst fractionation experiment (seeTable 1) starts off with a stibnite “China 3” from the ore collection (F1). The crushed ore was heated at 400°C in order to obtain Sb2O3. Then this roasted stibnite (F2) was
heated with C at 500°C during 4 h to Sb2O4 (F3a).
This was repeated with the same conditions for complete oxidation of the ore to Sb2O5 (F3b). A
frit composed of SiO2-Na2O-CaO (74.5:18:7.5),
which is close to average composition of Roman glass and at a eutectic point in the ternary phase dia-gram was created by heating the raw materials at 1200°C for 4 h (F4). The Sb2O5 (95:5) was added
to this frit andfired at 800°C for 4 h (F6a). Samples F1–4 were digested according to Procedure 2: the samples were weighed to 0.1 g after which 1.5 ml 12 M HCl was added. The beakers were closed and placed on the hotplate overnight followed by evapor-ation at 80°C on a hotplate. Then Aqua Regia (AR) was added, 0.8 ml 14 M HNO3 with 2.4 ml 12 M
HCl. The beakers were closed and heated again over-night at 90°C before they were opened to evaporate. Samples F5 and F6a were digested by the same procedure, with the additional step of adding 3 ml of 22 M HF with 2 ml of 14 M HNO3after
evapor-ating the Aqua Regia. The closed beakers were heated for two hours before they were opened to evaporate. Finally, 10 ml of 0.14 M HF was added to samples F1–F6a to redissolve the samples for Sb isolation before MC-ICP-MS analysis.
Fractionation experiment 2
The same experiment outline (Table 1) as in fraction-ation experiment 1 was followed, except for the adding of Sb. This time Sb2S3(F1) was added to the base glass
(F4) and heatedfirst at 800°C, then at 1200°C for twice 4 h (F6b). In this fractionation experiment Procedure 3,
Table 1.Fractionation experiment outline. Stibnite“China 3” was selected. The crushed ore was heated at 400°C in order to obtain Sb2O3. Then the roasted stibnite was heated with C at 500°C during 4 h. This was repeated with the same conditions for complete oxidation of the ore to Sb2O5. (F1–F3b were analysed by XRD). A frit composed of SiO2-Na2O-CaO (74.5:18:7.5), which is close to average composition of Roman glass and at a eutectic point in the ternary phase diagram is created by heating the raw materials at 1200°C for 4 h. Either a frit of Sb205(95:5) is added andfired at 800°C for 4 h (F6a) or Sb2S3is added to the base glass and heated first at 800°C, then at 1200°C for twice 4 h (F6b).
F1/L1 F2/L2 F3a F3b F4 F5/L5 F6a/L6a F6b/L6b Raw material Stibnite Roasted stibnite Oxidised stibnite Oxidised stibnite Raw materials glass Glass frit
Opacified Sb glass by adding Sb205
Opacified Sb glass by adding Sb2S3
T 400°C 500°C 500°C 1200°C 800°C 800–1200°C
offering total digestion for elemental analysis, was used. To closely monitor any change in oxidation state during the digestion procedure, every step of Procedure 3 was sampled. Firstly, all samples were weighed to 50– 70 mg in Savillex beakers to which 10 ml of AR was added (2.5 ml of 14 M HNO3+ 7.5 ml of 12 M HCl).
The samples were heated overnight at 120°C. Then, two times 1 ml and one time 2 ml was taken out (=4 ml in total) to respectively use for titration, ICP-OES analysis and Sb isotope analysis. The remaining 6 ml was evaporated at 150°C. 10 ml 22 M HF was added and heated overnight at the same temperature. Again, 4 ml was taken out for the analysis as described above. The remaining 6 ml was evaporated at 150°C and brought into solution again with 7 ml of 3% HNO3.
In this experiment stibnite “China 3” (F1) and an additional ore “Georgia 6” were digested with two different procedures in order to evaluate the influence between digestion procedures on the Sb isotopic com-position. Procedure 1 (Lobo et al.2012) and Procedure 3 were chosen. Procedure 1 goes as follows: a mixture of 4: 1: 1 HNO3: H2O2: HF was made (4 ml 14 M
HNO3+ 1 ml 10 M H2O2+ 1 ml 22 M HF), of which
1.5 ml was added to the samples and evaporated at 40°C. Then 2.5 ml of AR was added. The beakers were closed and put on the hotplate to heat overnight at 90°C. Afterwards, the beakers were opened and evaporated at the same temperature. The residue was redissolved in 10 ml of 0.14 M HF.
Fractionation experiment 3
This fractionation experiment copies the outline of experiment 1 (F1–F6a). A glass opacified by the adding of Sb2S3, as in experiment 2, was created as well (F6b).
These samples (F1–F6a/b) were digested with Pro-cedure 2, as described for experiment 1. Around 65 mg was taken from each samples, except for F3, of which 36 mg remained.
Additionally, these samples were also digested with Procedure 1, as described for experiment 2, and here labelled as L1 (stibnite) and L2 (roasted stibnite). The glass samples (L5 and L6a/b) had an additional addition of a mixture of 3 ml 22M HF and 2 ml 14M HNO3.
Methods XRD
The ores were crushed in an agate mortar and placed on the XRD holder using silica gel. Powders were measured using a Phillips PW1830 diffractometer with a Bragg/Brentanoθ–2θ setup and Cu Kα radiation at 45 kV and 30 mA. Angles from 5 to 75° 2θ were scanned with a step size of 0.02° 2θ and 1s per step. ConvX software was used forfile conversion, mineral identification was performed using DiffracPlus (EVA).
SEM-EDS/ESEM
A Hitachi 3500U Environmental Scanning Electron Microsocope (ESEM) at Cranfield University was used to obtain the major and minor element chemical composition of 10 jewellery samples, mainly beads found at the sites of Brili and Chalpiragorebi and 8 ores obtained in the Racha-Lechkumi region. The samples were mounted on carbon-coated stubs and analysed with 20 kV accelerating voltage, an intensity of 89μA, 30s live time and 10 mm working distance. The Octane Plus EDS detector (EDAX – AMTETEK Inc.) was used with a standardless approach. The measured intensities have been quantified using eZAF Smart Quant Results. Raw data were corrected with ZAF correction procedure to process for matrix effect. Multiple spots were analysed to calculate a mean value. The samples are given as elements with some exception presented as oxides. The results are presented with oxygen content included to indicate the corrosion condition of the metal.
ICP-OES
The major and minor elemental compositions of the ore collection samples were obtained by Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) analyses with a Varian 720-ES instrument (simultaneous ICP-OES with axially viewed plasma). Firstly, 42 ore samples were qualitatively analysed (only partial dissolution). Since they were similar in their major and minor elemental composition, sub-sequently only 19 ore samples were quantitatively ana-lysed. These 19 ore samples are discussed in the paper. 70 mg of ore was powdered. The powdered samples were subjected to digestion in Teflon beakers. Pro-cedure 3 was used to dissolve the samples (4 ml AR, 3 ml HF, 3 ml HNO3). The digest was transferred
into a larger tube (15 ml), and diluted to 8 ml using 1M HNO3. 5 ml of the 8 ml solution was transferred
back into Teflon beakers to be used for Sb isotope analysis (see below), while the remaining 3 ml was retained for chemical analysis by ICP-OES.
Exact instrumental setup can be found in Brems et al. (2012). The operating conditions and instrumen-tal parameters are summarised inTable 2.
MC-ICP-MS
Samples were weighed (around 50 mg, or less in case of a small sample) in Savillex beakers and digested according to Procedure 3 (4 ml AR, 3 ml HF), and eventually diluted in 3% HNO3for Sb concentration
check and isotope ratio measurement.
Neptune multi-collector inductively coupled plasma– mass spectrometer (MC–ICP–MS), equipped with a micro-flow PFA-50 Teflon nebulizer and a double spray chamber, consisting of a cyclonic and a
Scott-type sub-unit for sample introduction, and running in static multi-collection mode, at the University of Ghent (Belgium). It was operated with a constant flow of 3% HNO3at 100. The ion signal intensities of 121Sb and123Sb were detected simultaneously by
mul-tiple Faraday cups (see Table 3 for operating con-ditions) using 113In and 115In for the correction of the instrumental mass discrimination.
Revised Russell’s law using Indium (In) as internal standard was chosen as method of correction for instrumental mass discrimination (Baxter et al.2006), as it was shown by Lobo et al. (2013) that the latter method leads to improved precision and is reproduci-ble, in contrast to the sample-standard bracketing method. Sb isotopic results are reported in 123ε units. Theseε values express the isotopic composition of Sb as a relative difference or deviation with respect to an in-house standard (SPC Science), measured in a stan-dard-sample-standard bracketing sequence with a 1000 ppb Sb standard solution and 150 ppb In as an internal standard (Lobo et al.2012,2013,2014).
Results
Fractionation experiments
Figure 3 shows the results of the three fractionation experiments combined. A shift in isotopic composition of the different phases in the experiments is revealed,
Table 3.Operating conditions for the Faraday cups and data acquisition.
Thermo Scientific Neptune multi-collector ICP-mass spectrometer Cup configuration
L2 C H2 H3
113
In 115In 121Sb 123Sb Data acquisition parameters
Number of blocks 7 Number of cycles per block 5 Integration time (s) 4
Figure 3.Overview of the Sb isotopic composition of the fractionation experiments. Procedure 1 is denoted with triangles, Pro-cedure 2 with dots and ProPro-cedure 3 with diamonds. Experiment 1 is denoted in black colour, experiment 2 in pink and experiment 3 in red (if opacified by Sb2O5; F/L6a) and green (if opacified by Sb2S3/Sb2O3; F/L6b).
Table 2.ICP-OES instrumental settings.
Parameters VARIAN 720-ES: conditions used by all lines
Power kW 1.30
Plasmaflow L/min 15.0 Auxiliaryflow L/min 1.50 Nebulizerflow L/min 0.75 Replicate read time (s) 3.00 Instrumental stabilisation delay (s) 15 Sample introduction settings
Sample uptake delay (s) 60
Pump rate (rpm) 15 Rinse time (s) 5 Standards Standards quantitative % slope deviation 10 No of standards 15
except for the samples digested with Procedure 3 in experiment 2. The newly produced opacified glass of experiment 1 has a Sb isotope ratio which is around 3.0123ε units lower than the unprocessed ore. No sig-nificant change occurs between the unprocessed (F1) [Sb2S3] and roasted stibnite (F2) [Sb2O3]. After partial
oxidation (F3a) [Sb2O4] the isotopic values drop
around 1.5123ε. After adding the oxidised Sb (F3b) to the glass (F4), an opacified glass (F6a) is obtained. Its isotopic composition is again another 1.5–2.5 123ε
lower compared to the previous step. The same trend can be observed in both procedures (1 & 2) used in experiment 3, i.e. no shift occurs between the ore (F/ L1) and the roasted ore (F/L2), but does occur between the oxidised ore (F/L3) and the opacified glasses (F/ L6a/b). The shift is however larger for the samples digested with Procedure 2, i.e. between around 1.5 and 3.0 123ε value, while for the samples digested with Procedure 1 the shift ranges between barely a difference (0.0) and 2.2 123ε. For the opacified glasses
it is noted that more negative Sb isotopic values are found in the glasses opacified with the Sb oxide (F/ L6a) in comparison to the glasses opacified with stib-nite (F/L6b).
As can be seen in Figure 4, no fractionation between the different technological steps is observed in experiment 2. Neither is any significant fraction-ation occurring between the different steps of the digestion procedure (Procedure 3), nor when using HCl as first step. Figure 5 shows a discrepancy in
123ε Sb values between this Procedure 3 and
Pro-cedure 1. Samples digested with ProPro-cedure 3 have more negative values compared to samples digested with Procedure 1. A difference of 1.15 123ε is noted
for “Georgia 6” and 1.97 123ε for “China 3.” Since both samples were run without the isolation prep-aration, there is no impact at all possible of an event-ual on column fractionation. Likewise in experiment 3, the samples digested with Procedure 1 have a more negative Sb isotopic value than the ones digested with Procedure 2.
Ore collection
XRD and ICP-OES results
A summary of the results are shown inTables 4and5. Data can be found in Appendix I (XRD) and Appendix II (elemental composition). Table 4 shows that the sulfide, stibnite [Sb2S3], was identified the most. The
two other most common minerals identified in the col-lection were tetrahedrite-tennantite s.s. [(Cu,Fe)12Sb
4-S13] and boulangerite [Pb5Sb4S11]. Geochemical
analysis shows that these stibnites contain on average c. 75% Sb and c. 27% of S, indicating these stibnites occur without any significant content in any other major element. In only a few cases were As, Fe, Pb and Zn detected. Therefore, when used in ancient pro-cesses, stibnite would not influence in a major way an artefact’s composition other than for its Sb content. The tetrahedrite-tennantite s.s. has a different S – Sb ratio and an association with As, Fe, Zn and mainly Cu can be seen. In the case of the use of boulangerite, Pb is present in significant amounts (seeTable 5). Sb isotope analysis
The Sb isotope composition of stibnites in the ore col-lection (n = 63, including remeasured ores) could be determined with a good repeatability (internal pre-cision; the variation arising when all efforts are made to keep conditions constant by using the same method on the identical test items in the same laboratory by the same operator using the same equipment within short intervals of time) of 0.48123ε (2RSD), comparable to precisions reported in literature (e.g. similar to the
Table 4.Minerals identified by XRD.
Mineral
N =
90 Associated minerals Stibnite (Sb2S3) 64 Stibiconite, berthierite,
chalcopyrite, galena, sphalerite Tetrahedrite-tennantite s.s.
((Cu,Fe)12Sb4S13)
4 Galena, pyrite, sphalerite, chalcopyrite
Boulangerite [Pb5Sb4S11] 4 Falkmanite, galena, sphalerite
Other Sb-rich minerals 18
Figure 4.Isotopic composition of“China 3” (ore) during differ-ent digestion steps in experimdiffer-ent 2.
reported 0.4123ε by Lobo et al. (2012)). However, the reproducibility (obtained on replicates during the same run of the instrument) for high Sb materials (>10 wt% Sb) was bad, replicate measurements on stib-nites yielding a difference of up to 5123ε units. In
prac-tice, comparing values in the same batch was possible, but comparing between batches proved unreliable. Therefore, for high Sb content samples the preparation procedure for isotope measurements had to be re-developed. Recoveries were checked in elution profiles, which is the process of extracting one material from another by running a solvent, called an “eluent”, through loaded ion-exchange resins in columns to remove captured ions. After the solvent molecules dis-place the analyte (i.e. the isolation), the analyte can be carried out of the column for analysis.
A major loss of Sb was observed in elution profiles, especially after the second isolation step (see “elution 2e” inFigure 6). This loss can be ascribed to Sb that is retained in the resin on the column. Since geochem-ical analysis showed that not many associated elements are to be expected, chromatography was entirely
eliminated from the preparation procedure of the high-Sb archaeolgocial samples and from fractionation experiment 2, and high Sb content solutions were directly loaded for isotopic measurements.
Metallic Sb samples of the Racha– Lechkumi area
SEM-EDS and EPMA results
Six metal beads from the Brili site were analysed. Five beads were identified as metallic Sb and one ("Bead 1") as a Cu–Sn alloy (see Appendix II). On average the metallic Sb beads contain 69.0 wt% Sb. However, less corroded zones and original cores reveal that the original Sb content was much higher, e.g. 86.9 wt% in "bead 56". For Bead 1 an increase of Sn towards the outer zones due to the phenomenon of decuprification is observed (Figueiredo et al.2010; Oudbashi, Hasan-pour, and Davami 2016). Three metal samples from Chalpiragorebi were measured. SEM analysis reveal that they are made of metallic Sb, with contents up to 95.0 wt%, e.g. the metallic core in "CHPG 11". Cu is
Table 5.Average major and minor elements identified by ICP–OES.
Stibnite (n = 16)
Major (mean–median) (%) Traces (in few cases) (max) (%) Gangue material (in few cases) (max) (%)
Sb S Total As Fe Pb Zn Al Ca K
67.89 ± 16.88 25.13 ± 5.8 93.70 ± 47
74.0 27.1 101 0.699 0.536 0.518 0.149 2.06 1.69 0.101 Tetrahedrite-tennantite s.s. (n = 1)
Major (%) Traces (%) Gangue material (%) Total (%)
As Cu Fe Sb S Zn Bi Ca Mg
8.69 39.1 2.33 14.5 24.8 3.95 0.139 0.491 0.275 94.4 Boulangerite (n = 1)
Major (%) Traces (%) Gangue material (%) Total (%)
Pb Sb S As Fe Ca Mg Mn
6.49 25.4 5.60 1.00 0.647 0.509 0.287 0.107 40.2
attested only in the Brili beads, as 4 out of 5 samples yielded a Cu content of at least 1.00 wt%, and one even has almost 6 wt% Cu.
Sb isotope analysis
Sb isotope analysis has been carried out on the Geor-gian ore and metallic Sb samples in two runs. Data are listed in Appendix III.Figure 7shows the isotopic ratios for the samples analysed. Six of the Georgian ores plot above the 0.00 123ε value, while only one has a negative value (“G17”). On the contrary, all archaeological samples plot below the 0.00 123ε value except for Bead 10 (G32), which has a positive value.
Discussion
Fractionation during analysis (hypothesis 1) In Procedure 3 no shift in isotopic composition is observed. This can be explained by the fact that Sb is already present as Sb (+V) in the ore digests (as shown by titration) through the use of AR as strong oxidising agent in the digestion procedure. In this case, the digestion causes the transition in oxidation state, changing the Sb isotope composition of the orig-inal, roasted and oxidised ore. The risk of inducing fractionation in the digestion procedure is thus real (hypothesis 1).
Fractionation during pyrotechnological processes (hypothesis 2)
Our experiments revealed that fractionation does occur in making Sb opacified glass, altering the Sb isotopic
ratio to a lighter isotopic composition and hence more negative values.
Fractionation is known to happen in the transition of oxidation state (cfr. parallel to the redox reactions suggested by Ratié et al. (2016) for Ni isotopes) from +III to +V, which occurs when roasted stibnite (F2) is oxidised to Sb2O4(F3a) or Sb2O5(F3b) to add
as antimony source to the glass batch, or which occurs when Sb2S3/Sb2O3(+III) is added to the glass batch to
form Ca2Sb2O6/7(+V). It is put forward here that the
more Sb (+III) is present in thefinal glass, the closer the Sb isotopic value will be to the roasted and/or unprocessed ore. The more Sb (+V) is present, the more negative the Sb isotopic ratio will be due to fractionation.
First, in the process of roasting stibnite no difference in Sb isotopic ratio is seen (<SD) (Table 6) between unprocessed and roasted ore, thus no fractionation occurs in this procedure. In oxidised ore, no significant shift (i.e. around SD) is observed when Sb2O3 is
obtained, but a difference of 1.39 123ε is seen when
Sb2O4is obtained, due to the mixed valence in the
lat-ter. The opposite trends (i.e. higher Sb isotope ratio compared to the roasted ore) for the oxidised ores [Sb2O4, cervantite] in Procedure 2 of experiment 1
can be explained by the uncontrolled parameters in this experiment (seeFigure 3), such as oxygen fugacity or partial pressure during heating in the oven. Hence, the ratio of Sb3+/Sb5+ in these cervantites will vary, and oxidised ores with more Sb(+III) still present will have a Sb isotopic ratio closer to the roasted ore. If more Sb (+V) will be present, then the Sb isotopic
ratio will be more negative and hence closer to the iso-topic composition of the opacified glasses.
Finally, in the process of making an opacified glass, and thus theoretically going completely to a + V valence, differentiation is seen among the different digestion procedures. While for Procedure 3 no frac-tionation seems to occur (0.13 123ε value <SD), Pro-cedure 2 shows that fractionation from a stibnite to a Sb oxide opacified glass gives a shift in Sb isotopic value of around 2.5 123ε value. Interestingly, a lower shift is observed when stibnite was used directly in the opacification process (1.63 123ε). For Procedure
1, a shift from stibnite to the Sb oxide opacified glass of 2.11 123ε is noted, whereas for the glass opa-cified with the stibnite this is only 0.55 123ε value. In
this case, it seems that not all Sb (+III) was oxidised to Sb (+V) during the in-situ opacification process. Hence, the oxidation state of the Sb in the glass matrix stays mainly the same as the Sb source. These findings are in line with the observations made by Lahlil et al. (2008), who states that if Sb (+III) is the predominant species in the glass matrix and if it is known that the glass has been made under conditions below 1200°C, the sources of Sb poten-tially used are compounds of the type Sb2O3 or
Sb2S3 (Lahlil et al. 2008). From our experiments it
can be concluded that the transition in oxidation stage from +III to +V during the technological pro-cess of using stibnite in glass making leads to a drop in the Sb isotopic ratio of c. 2.5 123ε. This frac-tionation factor (0.25‰) can be used to interpret any further Sb isotopic data from archaeological data.
Therefore, the assumption that there is a direct relationship between mineral raw minerals and the artefacts, so that the geological signature is not trans-formed physically or chemically during the manufac-turing process (Ixer1999), is not fulfilled. The goal of
using Sb isotopic compositions for the determination of provenance can only be aimed for when Sb stays as Sb (+III) (e.g. Lobo et al.2012) or when a fraction-ation factor is taken into account. Moreover, since this fractionation happens during technological processing, and since we can control the digestion procedures used in measuring Sb isotopes, this method can be of use with regard to technological issues in antiquity. Since different technological processes can determine whether Sb enters the glass batch in Sb(+III) or as Sb (+V), it might give us some insight in the production
of antimonates. Previous study by Degryse et al. (2015) (Figure 8) suggested that the Sb used in LBA glass making has a distinct source from later Sb used in Greco-Roman times as a decoulouriser, and that the LBA glass has a signature similar only to remains of early Sb metallurgy, developed in the Caucasus. Met-allic or mineral Sb was used to opacify glass, and fac-tories in Mesopotamia and Egypt seem to have used the same source of the material (Degryse et al.2015). It can indeed be observed (see the red line onFigure 8) that the LBA glass has a negative Sb isotopic compo-sition, while the Roman glasses show higher Sb isotopic compositions. Since fractionation is occurring in the oxidising phase, this difference in isotopic composition can no longer only be explained by a different Sb source, but also by a different technological process.
Using the principle of Occam’s razor, a schematic overview for the mass-dependent Sb fractionation as partially deduced experimentally is presented inFigure 9. Since the most straightforward procedure to make metallic Sb starts with roasting stibnite (+III to +III/ +V) and consequently reducing it in closed moulds, the same fractionation factor for metallic Sb beads is proposed, as it would induce redox reactions as well. However, the oxidation state of the metallic beads remains difficult to predict. The roasting of the stibnite renders either Sb2O3(+III) or Sb2O4(+III/+V). While a
oxidation state of a metal is always (0), this is not necessarily the case when we are dealing with artefacts that are not a full 100% metallic Sb. In that case, redu-cing might turn it (partially) into +III and/or Sb(0).
Case study: provenancing the metallic Sb beads The elemental composition of the ores analysed, leads us to conclude that the use of stibnite in ancient indus-tries would not contribute many elements to the end product other than Sb and S. Based on their ubiquity in ore deposits, only the tetrahedrite-tennantite s.s. and boulangerite are likely other ore minerals to be used. When a tetrahedrite-tennantite s.s. is used, the presence of As, Fe, Zn and mainly Cu as accompanying elements is expected. In the case of boulangerite, Pb would be a significant associated element. Given the interesting geological and archaeological background of the Caucasus, such patterns were investigated in the composition of the metallic Sb samples. Afirst indi-cation for the use of a different antimony source and/or technology can be suggested: in the Brili beads the
Table 6.Average123ε value of the stibnites and roasted ores for the different digestion procedures in the different experiments.
association of Sb and Cu hints at the use of tetrahe-drite-tennantite s.s. ore, whereas the lack of associated elements for the Chalpiragorebi beads suggests stibnite instead.
Using the scheme for fractionation correction and in view of the uniqueness of the Caucasian deposits and the strong archaeological tie between these ores and the very existence of Sb metallurgy, it is suggested that the metallic Sb beads of Brili and Chalpiragorebi
are consistent with the local Sb sources of the Racha-Lechkumi region.
In light of the redox processes and fractionation, the negative value for “G17” is interesting (Figure 7). “G17” was originally labelled as ore, found in grave 17 on the site of Brili. However, its XRD shows an amorphous phase and an unattributable antimony peak (Sb, Sb2O3, Sb2S3…). Chemically, “G17” contains
57.6 wt% Sb, 2.5 wt% Fe, and 1.3 wt% As. Therefore, it is most likely evidence of the smelting practice, consid-ering the amorphous phase and the high amounts of Sb, Fe and As. In that case, the negative isotopic value for“G17” can also be explained by the fraction-ation occurring in the redox process during the metal-lurgical processing, which seems to follow the same pattern seen in the glass making process. An alternative explanation, based on the redox process, to understand the negative value is that the ore got oxidised under burial circumstances similarly to the valentinite beads at Selvicciola (Italy). Pallecchi, Pecchioli, and Tocci (2002) consider the valentinite nature of the beads as the result of postdepositional processes (oxidation of Sb) due to their small size.
However, the experiments showed that the digestion Procedure 3 also causes a shift of +III to +V in the stib-nites. Since these samples have been digested with this procedure, it should be expected that the stibnites have shifted in oxidation state and hence give Sb isotopic values that are 0.25‰ lower than if they would still
Figure 8.Blue arrows point out the difference between the 123ε values obtained for Spain 4, Italy 3 and the Jerablus bead and the values obtained for these identical samples by Lobo et al. (2012); orange shapes show the correction with the fractionation factor between the unprocessed ores of the Racha region and the metallic beads of Brili and Chalpiragorebi, and the red line indicates the observation made by Degryse et al. (2015) that the Egyptian and Mesopotamian glasses have a lower value than the Roman glasses. *Data obtained from Degryse et al. (2015).
be in their original state +III. It remains unclear whether the still highly positive values of the Rachan ores mean that no redox reactions have occurred during digestion or if/whether the metallic beads were affected by this digestion procedure.
Conclusion
Sb as a raw material was frequently used, both as an opacifier and decolouriser in ancient glasses, as well as an alloying element or in its metallic phase in metal production. Despite this ubiquity, antimony pro-duction has only rarely been studied (for medieval times and later, see Siebenschock1996) and questions concerning its provenance and origin remain unanswered.
While this paper has presented evidence from a rela-tively limited number of specific laboratory exper-iments, the results have general significance for the interpretation of Sb isotope compositions as a whole. Firstly, the experiments revealed significant fraction-ation of stable Sb isotopes during the conversion of stibnite (Sb+III) into Sb oxide (Sb+V). The range of frac-tionation appears to be around123Δ Sb = 0.25‰. Since the induced isotopic shifts can be explained by redox processes, these isotopes can be largely corrected for and being used as a fingerprint with the relation to the source of the element. Hence, for provenancing purposes, fractionation can be taken into account and corrected for when looking for a consistent match with a source, as the Caucasian case study has demon-strated. Secondly, Sb isotopes can be used to differen-tiate between natural and anthropogenic sources, potentially revealing information about ancient anti-mony smelting and production techniques. This includes fractionation during technological processing. These results also highlight the potential value of the method to distinguish different technological processes in ancient pyroindustries. Extended fractionation experiments in future research should evaluate the influence of fractionation on Sb-rich metal artefacts (reduction) to model this potential shift quantitatively. Future research should link these results with archaeo-metallurgical and ancient glass making processes, and look into the Caucasus. This region is of interest, especially the Racha-Lechkumi district in Georgia. Elemental and isotopic data of LBA metallic Sb objects and stibnites from this region were obtained. They suggested that both stibnite and tetrahedrite-tennantite s.s. could have been used as a raw material. Based on the fractionation scheme the Sb isotopic data of the archaeological artefacts are consistent with the local Sb sources. Unfortunately, the digestion procedure used complicates the picture and more research is needed on the link between fractionation and reducing antimony before confident claims about provenance of the metallic Sb jewellery found can be made.
Acknowledgements
Prof Dr Davit Lordkipanidze, the Director of the Georgian National Museum (GNM) and the Royal Belgian Institute Natural Sciences (KBIN) are acknowledged for the per-mission to sample, respectively, the metallic beads and stib-nite ores. We are thankful to Givi Inanishvilii and Gela Gobejeshvili, members of the previous expedition to the study area, under the auspices of the Academy of Science. I, Peter, would also like to thank the people of Gona who guided me safely back from the expedition. We are grateful to Elvira Vassilieva for help with the sample preparation and ICP–OES analyses. Kris Lauwers of Ghent University is thanked for running the Neptune measurements. Dr Den-nis Braekmans is warmly acknowledged for providing the ESEM data.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This work was supported by FWO Vlaanderen [grant num-ber G0C4315N].
Notes on contributors
Sarah Dillisis a Ph.D. student at the KU Leuven (Belgium) working on the raw material antimony, in ancient times, focussing on provenance determination of this raw material used in vitreous and metal making by the application of Sb isotope analysis.
Alicia van Ham-Meert, having completed an M.Sc. in chemical and materials engineering and in archaeological science, her main doctoral research has focused on non-inva-sive chemical and isotopic analysis (KU Leuven). Now within the Timber project she focuses on Sr isotopic analysis for tracing medieval timber trade.
Peter Leemingis a graduate from the Department of Archae-ology and honorary research fellow at the University of Exe-ter (UK). He has been working in several projects in conjunction with the National Museum of Georgia. Andrew Shortlandis professor leading the Cranfield Foren-sic Institute (UK). His expertise is the identification and interpretation of material culture from the ancient and his-torical worlds. He is particularly involved in the analysis of glass, glaze and ceramics of a wide range of dates.
Gela Gobejishviliis an anthropologist and archaeologist, and the curator of the Brili archive (Georgian National Museum). Mikheil Abramishvili has been the director of the Tbilisi Archaeological Museum (now part of the Georgian National Museum) for the last 20 years. His research interests com-prise the development of ancient metallurgy, trade relations and religious beliefs in the Old World.
ORCID
Alicia Van Ham-Meert http://orcid.org/0000-0001-6796-3141
References
Asael, D., A. Matthews, M. Bar-Matthews, and L. Halicz.
2007. “Copper Isotope Fractionation in Sedimentary Copper Mineralization (Timna Valley, Israel).” Chemical Geology 243 (3–4): 238–254. doi:10.1016/j.chemgeo. 2007.06.007.
Baxter, D. C., I. Rodushkin, E. Engström, and D. Malinovsky.
2006. “Revised Exponential Model for Mass Bias Correction Using an Internal Standard for Isotope Abundance Ratio Measurements by Multi-Collector Inductively Coupled Plasma Mass Spectrometry.” Journal of Analytical Atomic Spectrometry 21 (4): 427– 430.doi:10.1039/b517457k.
Belli, O., and E. Konyar. 2001. “Excavations at Van-Yoncatepe Fortress and Necropolis.” Journal of the Institute of Tel Aviv University 28 (2): 169–212. doi:10. 1179/tav.2001.2001.2.169.
Berger, D., E. Figueiredo, G. Brügmann, and E. Pernicka.
2018. “Tin Isotope Fractionation During Experimental Cassiterite Smelting and Its Implication for Tracing the Tin Sources of Prehistoric Metal Artefacts.” Journal of Archaeological Science 92: 73–86. doi:10.1016/j.jas.2018. 02.006.
Brems, D. 2012. “Mineralogy and Geochemistry of Mediterranean Sand Deposits as a Raw Material for Roman Natron Glass Production.” Journal of Archaeological Science 39.
Brems, D., and P. Degryse.2014.“Trace Element Analysis in Provenancing Roman Glass-Making*.” Archaeometry 56 (September 2013): 116–136.doi:10.1111/arcm.12063. Brems, D., P. Degryse, F. Hasendoncks, D. Gimeno, A.
Silvestri, E. Vassilieva, S. Luypaers, and J. Honings.
2012. “Western Mediterranean Sand Deposits as a Raw Material for Roman Glass Production.” Journal of Archaeological Science 39 (9): 2897–2907. doi:10.1016/j. jas.2012.03.009.
Cambi, L., and F. Cremascoli.1957.“Sul Metallo Dei Bottoni Della Tomba Preistorica Di Monte Bradoni Presso Volterra.” RendIstLomb 91: 371–377.
Carmi, I., C. Epstein, and D. Segal.1995.“Radiocarbon Dates from Chalcolithic Sites in the Golan.” Atiqot 26: 207–209. Chernykh, E. N.1992. Ancient Metallurgy in the USSR: The Early Metal Age. Cambridge: Cambridge University Press. Cui, J., X. Wu, and B. Huang. 2011. “Chemical and Lead Isotope Analysis of Some Lead-Barium Glass Wares from the Warring States Period, Unearthed from Chu Tombs in Changde City, Hunan Province, China.” Journal of Archaeological Science 38 (7): 1671–1679.
doi:10.1016/j.jas.2011.02.034.
Degryse, P., L. Lobo, A. Shortland, F. Vanhaecke, A. Blomme, J. Painter, D. Gimeno, et al. 2015. “Isotopic Investigation Into the Raw Materials of Late Bronze Age Glass Making.” Journal of Archaeological Science 62: 153–160.doi:10.1016/j.jas.2015.08.004.
De Marinis, R. C.2006.“Aspetti Della Metallurgia Dell’età Del Rame e Dell’antica Età Del Bronzo.” Toscana. Rivista Di Scienze Preistoriche 56: 211–272.
Dolfini, A. 2010. “The Origins of Metallurgy in Central Italy.” Antiquity 84 (325): 707–723.
Dolfini, A. 2014. “Early Metallurgy in the Central Mediterranean.” In Archaeometallurgy in Global
Perspective, edited by B. W. Roberts and C. P. Thornton, 473–506. Springer.
Figueiredo, E., R. J. C. Silva, F. M. Braz Fernandes, and M. F. Araújo.2010. “Some Long Term Corrosion Patterns in Archaeological Metal Artefacts.” Materials Science Forum 636–637: 1030–1035. www.scientific.net/MSF. 636-637.1030.
Fleming, S. J., S. K. Nash, and C. P. Swann. 2011. “The Archaeometallurgy of Period IVB Bronzes at Hasanlu.” In Peoples and Crafts in Period IVB at Hasanlu, Iran, edi-ted by M. de Schauensee, 103–134. Philadelphia: Penn Press.
Freestone, I. C., and C. P. Stapleton.2015. “Composition, Technology and Production of Coloured Glasses From Romman Mosaic Vessels.” In Glass of the Roman World, 61–76. Oxford: Oxbow Books.
Gobejishvili, G. F. 1952. არქეოლოგური გათხრები საბჭოთა საქართველოში(თბილისი). [Archaeological Excavations in the Soviet Era] (Tbilisi).
Gobejishvili, G. F. 1977. “‘ბრილი’ ქართული საბჭოთა ენციკლოპედია (თბილისი). [‘Brili’ – Article in Georgian SSR Encyclopedia].”
Golden, J.2014.“Who Dunnit? New Clues Concerning the Development of Chalcolithic Metal Technology.” In Archaeometallurgy in Global Perspective: Methods and Syntheses, edited by Benjamin W. Roberts and Christopher Peter Thornton, 559–578. New York, NY: Springer New York.
Grazzi, F., P. Pallecchi, P. Petitti, A. Scherillo, and M. Zoppi.
2012. “Non-Invasive Quantitative Phase Analysis and Microstructural Properties of an Iron Fragment Retrieved in the Copper-Age Selvicciola Necropolis in Southern Tuscia.” Journal of Analytical Atomic Spectrometry 27 (2): 293–298.doi:10.1039/C1JA10184F. Hauptmann, A., and I. Gambaschidze. 2001. “Antimon –
Eine Metallurgische Besonderheit Aus Dem Kaukasus.” In Georgien – Schätze Aus Dem Land Des Goldenen Vlies. Ausstellungskatalog, 150–155. Bochum: Deutsches Bergbau Museum.
Inanishvili, G., B. Maisuradze, and G. Gobejishvili. 2010. Mining and Metallurgical Activity in Ancient Georgia (3rd–1st Millennium B.C.). Tbilisi: Artlines.
Ixer, R. A.1999.“The Role of Ore Geology and Ores in the Archaeological Provenancing of Metals.” In Metals in Antiquity. BAR International Series 792, edited by S. M. M. Young, A. M. Pollard, P. Budd, and R. A. Ixer, 43– 52. Oxford: Archaeopress.
Kekelia, S. A., M. A. Kekelia, S. I. Kuloshvili, N. G. Sadradze, N. E. Gagnidze, V. Z. Yaroshevich, G. G. Asatiani, J. L. Doebrich, R. J. Goldfarb, and E. E. Marsh. 2008. “Gold Deposits and Occurrences of the Greater Caucasus, Georgia Republic: Their Genesis and Prospecting Criteria.” Ore Geology Reviews 34 (3): 369–386. doi:10. 1016/j.oregeorev.2008.04.003.
Lahlil, S., I. Biron, L. Galoisy, and G. Morin. 2008. “Rediscovering Ancient Glass Technologies Through the Examination of Opacifier Crystals.” Applied Physics A 92 (July): 109–116.doi:10.1007/s00339-008-4456-8.
Lindsay, I., and A. T. Smith.2006.“A History of Archaeology in the Republic of Armenia.” Journal of Field Archaeology 31 (2): 165–184.doi:10.1179/009346906791072016. Lobo, L., P. Degryse, A. Shortland, K. Eremin, and F.
Vanhaecke. 2014. “Copper and Antimony Isotopic Analysis via Multi-Collector ICP-Mass Spectrometry for Provenancing Ancient Glass.” Journal of Analytical Atomic Spectrometry 29 (1): 58–64. doi:10.1039/
Lobo, L., P. Degryse, A. Shortland, and F. Vanhaecke.2013. “Isotopic Analysis of Antimony Using Multi-Collector ICP-Mass Spectrometry for Provenance Determination of Roman Glass.” Journal of Analytical Atomic Spectrometry 28 (8): 1213–1219.doi:10.1039/c3ja50018g. Lobo, L., V. Devulder, P. Degryse, and F. Vanhaecke.2012.
“Investigation of Natural Isotopic Variation of Sb in Stibnite Ores via Multi-Collector ICP-Mass Spectrometry – Perspectives for Sb Isotopic Analysis of Roman Glass.” Journal of Analytical Atomic Spectrometry 27 (8): 1304.doi:10.1039/c2ja30062a. Maisuradze, B., and G. Gobejishvili.2001.“Alter Bergbau in
Ratscha.” In Georgien Schätze Aus Dem Land Des Goldenen Vlies. Ausstellungskatalog, edited by I. Gambaschidze, A. Hauptmann, R. Slotta, and Ü. Yalçin, 130–135. Bochum: Deutsches Bergbau Museum. Mass, J. L., M. T. Wypyski, and R. E. Stone.2002.“Malkata
and Lisht Glassmaking Technologies: Towards a Specific Link Between Second Millennium BC Metallurgists and Glassmakers.” Archaeometry 44 (1): 67–82.doi:10.1111/ 1475-4754.00043.
Mathur, R., S. Titley, F. Barra, S. Brantley, M. Wilson, A. Phillips, F. Munizaga, V. Maksaev, J. Vervoort, and G. Hart. 2009. “Exploration Potential of Cu Isotope Fractionation in Porphyry Copper Deposits.” Journal of Geochemical Exploration 102 (1): 1–6. doi:10.1016/j. gexplo.2008.09.004.
Mattielli, N., J. C.J. Petit, K. Deboudt, P. Flament, E. Perdrix, A. Taillez, J. Rimetz-Planchon, and D. Weis. 2009. “Zn Isotope Study of Atmospheric Emissions and Dry Depositions Within a 5 Km Radius of a Pb-Zn Refinery.” Atmospheric Environment 43 (6): 1265–1272.
doi:10.1016/j.atmosenv.2008.11.030.
Meliksetian, K., S. Kraus, E. Pernicka, P. Avetissyan, S. Devejian, and L. Petrosyan. 2011. “Metallurgy of Prehistoric Armenia.” Der Anschnitt 24 (180): 201–210. Meliksetian, K., and E. Pernicka. 2003. “Geochemical
Characterisation of Armenian Early Bronze Age Metal Artefacts and Their Relation to Copper Ores.” Proceeding of Archaeometallurgy in Europe 56 (1): 41–58. Moorey, P. R. S. 1999. Ancient Mesopotamia Materials and Industries: The Archaeological Evidence. Oxford: Clarendon. Oudbashi, O., A. Hasanpour, and P. Davami. 2016. “Investigation on Corrosion Stratigraphy and Morphology in Some Iron Age Bronze Alloys Vessels by OM, XRD and SEM–EDS Methods.” Applied Physics A: Materials Science and Processing 122 (4): 1–11.
Pallecchi, P., R. Pecchioli, and A. M. Tocci. 2002. “La Necropoli Eneolitica Della Selvicciola (Ischia Di Castro – VT): I Vaghi Della Tomba 23.” PPE Atti 539–543. Pantskhava, L., B. Maisuradze, and G. Gobejishvili. 1939.
“ბრილის სამარტოვანზე 1939 წელს გათხრილი No.12 სამარხის დათარიღებისათვის.” ძიებანი 8): 39–45.
Petitti, P., C. Persiani, and P. Pallecchi. 2012. “Reperti Metallici Dalla Necropoli Della Selvicciola (Ischia Di Castro – Viterbo).” Atti Della XLIII Riunione Scientifica IIPP, “L’età Del Rame in Italia”, Bologna, 26-29 Novembre 2008 1: 187–194.
Petrie, W. M. F.1891. Illahun, Kahun and Gurob. London: David Nutt.
Pike, A. W. G.2002.“Analysis of Caucasian Metalwork – the Use of Antimonal, Arsenical and Tin Bronze in the Late Bronze Age. Ancient Caucasian and Related Materials in the British Museum.” British Museum Occasional Paper 121: 89–92.
Ratié, G., C. Quantin, D. Jouvin, D. Calmels, V. Ettler, Y. Sivry, L. Cruz Vieira, E. Ponzevera, and J. Garnier.2016. “Nickel Isotope Fractionation During Laterite Ni Ore Smelting and Refining: Implications for Tracing the Sources of Ni in Smelter-Affected Soils.” Applied Geochemistry 64: 136–145. doi:10.1016/j.apgeochem. 2015.09.005.
Sagona, A. 2017. The Archaeology of the Caucausus. From Earliest Settlements to the Iron Age. New York: Cambrige World Archaeology.
Sayre, E. V. 1963. “The Intentional Use of Antimony and Manganese in Ancient Glasses.” In Advances in Glass Technology, Part 2, edited by F. R. Matson and G. Rindone, 263–282. New York: Plenum Press.
Schreurs, J. W. H., and R. Brill. 1984. “Iron and Sulfur Related Colors in Ancient Glasses.” Archaeometry 26: 199–209.
Shalev, S., and J. P. Northover.1993.“The Metallurgy of the Nahal Mishmar Hoard Reconsidered.” Archaeometry 35 (1): 35–47.doi:10.1111/j.1475-4754.1993.tb01022.x. Shortland, A. J.2002.“The Use and Origin of Antimonate
Colorants in Early Egyptian Glass*.” Archaeometry 44 (4): 517–530.
Shortland, A. J.2012. Lapis Lazuli from the Kiln: Glass and Glassmaking the Late Bronze Age. Studies in Archaeological Sciences. Vol. 2. Leuven: Leuven University Press.
Siebenschock, M. 1996. “Archäometallurgische Untersuchungen Zur Spätmittelalterlichen Und Frühneuzeitlichen Verhüttung von Antimonerzen Bei Sulzburg Im Südschwarzwald” 84/85 (Ber. Naturf. Ges. Freiburg i. Br.): 5–26
Stapleton, C. P. 2011. “Glass and Glaze Analysis and Technology from Hasanlu, Period IVB.” In Peoples and Crafts in Period IVB at Hasanlu, Iran, 87–102
Tadmor, M., D. Kedem, F. Begemann, A. Hauptmann, E. Pernicka, and S. Schmitt-Strecker. 1995. “The Nahal Mishmar Hoard from the Judean Desert : Technology, Composition, and Provenance.” Atiqot XXVII: 95–148.
Tite, M. S.2002.“The Development of Vitreous Materials in the Ancient Near East and Egypt. (BL: DN052656)” U155609 (8): 585–93. http://search.proquest.com/ docview/301571013?accountid=9883%5Cnhttp://elibrary. cf.ac.uk/sfx?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/ fmt:kev:mtx:dissertation&genre=dissertations+& +theses&sid=ProQ:ProQuest+Dissertations+&+Theses: +UK+&+Ireland&atitle=&title=T.
Wombacher, F., M. Rehkämper, K. Mezger, and C. Münker.
2003. “Stable Isotope Compositions of Cadmium in Geological Materials and Meteorites Determined by Multiple-Collector ICPMS.” Geochimica et Cosmochimica Acta 67 (23): 4639–4654. doi:10.1016/ S0016-7037(03)00389-2.